Hello everyone, Our studio recently moved from a well known engine to another : Unity. This decision is motivated by Unity's potential in performance saving and HD render pipeline development.
In our game, the camera displays a large scene with thousands of trees at the time ; we then decided to use an impostoring system and amplify's announcement made its effect on us.
After a few days trying different possibilities, i'd like to share my feeling about this and hopely get feedback from you guys on the issues i experienced.
First, A.Impostors is indeed really intuitive to use ; one click bake is reel. It has a few parameters to play with, but in the end you'll always get your impostors easily ; which is great.
Now let's get into the weird stuff. For our test we planned to use 5 tree species; create impostors from highpoly models and then duplicate them in our scene with different rotation and scale values to fake variations. We figured out that a same impostor, duplicated with different scales still conserves same parameters for both of them; resulting in weird atifacts, especialy when the camera is moving. I suppose duplicating them doesnt create shader instancies, which gives an answer. From this point we decided not to use different scale for same species and the first issue was resolved.
And here happened the second issue. From a certain camera angle, our impostors are flickering (see GIF). https://gfycat.com/ForsakenEarnestAllensbigearedbat I tryed disabling Post process/volume profile, tested the channels one by one, changed the lighting ... nothing changes. Do you have any idea on whats happening ?
Hello, thank you for getting in touch and for your support!
It's very hard to tell by just the gif itself, we've come across a similar behavior in another project that had a moving light source, which is still being investigated by the developer.
Which Unity, HD RP and AI versions are you currently using? Are you using standard shaders or custom ones in your objects?
Can you replicate the issue in a new project and send us a simple sample for us to debug on our end using the same data as you? Feel free to send it through firstname.lastname@example.org with the link to this post and we'll look into it as soon as possible.